Detection of Sinus Bradycardia with Electrocardiogram using Machine Learning Techniques

Cardiovascular Diseases (CVDs) are a group of diseases related to the heart or blood vessels. The diagnosis and subsequent treatment of CVDs poses significant challenges and necessitates accurate detection methods. In this study, we implement seven different machine learning techniques to determine the most suitable approach for detecting sinus bradycardia, a manifestation of CVD, using Electrocardiogram (ECG) signals. We present a comparative assessment in order to show diagnostic superiority of our proposed artificial neural network (ANN) model on a publicly available dataset using only 94 records. The proposed light-weight ANN model emerges as the most effective method with 95.2% accuracy in circumstances where data is limited. Plausible accuracies of over 90% are shown by KNN (K=3), Support Vector Machines and Random Forest which underscore the potential of machine learning algorithms in facilitating clinicians for effective diagnosis of sinus bradycardia.